Re: [casper] progdev fail
Hi all, Sent info to jack. If anyone wants direct info I will send to requestors to avoid cluttering the mailing list with attachments. Regards mm. From: Jonathan Weintroub [jweintr...@cfa.harvard.edu] Sent: 15 December 2016 07:44 PM To: Jack Hickish Cc: Mike Movius; Marc Welz; Franco; Casper Lists Subject: Re: [casper] progdev fail This may be a partially imperfect recollection in the details: Laura Vertatschitsch worked on a DDR3 to CPU interface at SAO. However, after substantial effort, she still saw irreducible bit errors at some low rate, detectable with write-readback testing for extended time periods. I vaguely recall the difficulty in this design arises in crossing the clock domain boundary to the PPC perhaps causing metastability and requiring double buffering or similar. In the end we gave up on this approach, and figured out how to dump our data via the FPGA and a 10 GigE port to a host computer with a suitable NIC. This workaround is still successfully in use, though arguably the PPC 1 Gig Ethernet would have been a more convenient solution at the time. Laura may be able to fill in the details. Interested to see Mike worked on this, and assuming his interface is entirely error free, we’d love to see it on the repo. Jonathan On Dec 15, 2016, at 10:39 AM, Jack Hickish <jackhick...@gmail.com<mailto:jackhick...@gmail.com>> wrote: Hi Mike, That would be great, thanks! Jack On Thu, 15 Dec 2016 at 15:11 Mike Movius <mi...@reutech.co.za<mailto:mi...@reutech.co.za>> wrote: Hi, I have managed to modify the ddr3 pcore for the roach2 to include the CPU interface. I am working from a cloned ska-as mlib_devel repo. I sent the changes/additions to Wesley New for vetting/committing to their repo but I guess he hasn't done it yet. I can send the info to those interested. Regards, mm. From: casper-boun...@lists.berkeley.edu<mailto:casper-boun...@lists.berkeley.edu> [casper-boun...@lists.berkeley.edu<mailto:casper-boun...@lists.berkeley.edu>] on behalf of Jack Hickish [jackhick...@gmail.com<mailto:jackhick...@gmail.com>] Sent: 15 December 2016 12:32 PM To: Marc Welz; Franco Cc: Casper Lists Subject: Re: [casper] progdev fail More broadly speaking, has anyone ever used the DDR powerPC interface on ROACH2? Is it actually know to work with code as-provided in the mlib_devel / katcp repos? Franco -- perhaps this project -- https://github.com/argonnexraydetector/RoachFirmPy<https://github.com/argonnexraydetector/RoachFirmPy><https://github.com/argonnexraydetector/RoachFirmPy<https://github.com/argonnexraydetector/RoachFirmPy>> -- which I found linked on the DDR page of the CASPER wiki, might be interesting to you. Cheers Jack On Thu, 15 Dec 2016 at 07:31 Marc Welz <m...@ska.ac.za<mailto:m...@ska.ac.za><mailto:m...@ska.ac.za<mailto:m...@ska.ac.za>>> wrote: The fpga<->ppc interface is limited in size - if you want to see huge memory areas, you will need some sort of windowing/offset scheme regards marc On Wed, Dec 14, 2016 at 1:43 PM, Franco <francocuro...@gmail.com<mailto:francocuro...@gmail.com><mailto:francocuro...@gmail.com<mailto:francocuro...@gmail.com>>> wrote: > Doing some more testing, I realize the progdev works when I remove the DRAM > block from the model. What could this mean?? > > Franco > > > > On 14/12/16 10:13, Franco wrote: >> >> Dear Casperites, >> >> I'm trying to progdev a model to ROACH2 and I get the following error: >> >> ?progdev cntr_dram2_2016_Dec_13_1558.bof >> #log info 1019840342915 raw >> attempting\_to\_program\_cntr_dram2_2016_Dec_13_1558.bof >> #log info 1019840343002 raw >> attempting\_to\_program\_bitstream\_of\_19586188\_bytes\_to\_device\_/dev/roach/config >> #fpga loaded >> #log warn 1019840343887 raw >> requesting\_to\_map\_a\_rather\_large\_area\_of\_0x8000 >> #log error 1019840343887 raw >> unable\_to\_map\_file\_/dev/roach/mem:\_Cannot\_allocate\_memory >> #log error 1019840343890 raw >> unable\_to\_program\_bit\_stream\_from\_cntr_dram2_2016_Dec_13_1558.bof >> !progdev fail >> >> Anyone has any idea what's the problem? I can program other models fine, >> but this is the first model I compiled using the SKA fork library. >> >> Thanks, >> >> Franco >> > > Please consider the environment before printing this e-mail View the Reutech Radar System online disclaimer at http://www.rrs.co.za/links/E-maildisclaimer.asp<http://www.rrs.co.za/links/E-maildisclaimer.asp> <116121517112301343.gif><116121517112301343.gif> Please consider the environment before printing this e-mail. DISCLAIMER: This message is inten
Re: [casper] progdev fail
This may be a partially imperfect recollection in the details: Laura Vertatschitsch worked on a DDR3 to CPU interface at SAO. However, after substantial effort, she still saw irreducible bit errors at some low rate, detectable with write-readback testing for extended time periods. I vaguely recall the difficulty in this design arises in crossing the clock domain boundary to the PPC perhaps causing metastability and requiring double buffering or similar. In the end we gave up on this approach, and figured out how to dump our data via the FPGA and a 10 GigE port to a host computer with a suitable NIC. This workaround is still successfully in use, though arguably the PPC 1 Gig Ethernet would have been a more convenient solution at the time. Laura may be able to fill in the details. Interested to see Mike worked on this, and assuming his interface is entirely error free, we’d love to see it on the repo. Jonathan > On Dec 15, 2016, at 10:39 AM, Jack Hickish <jackhick...@gmail.com> wrote: > > Hi Mike, > > That would be great, thanks! > Jack > > On Thu, 15 Dec 2016 at 15:11 Mike Movius <mi...@reutech.co.za > <mailto:mi...@reutech.co.za>> wrote: > > Hi, > I have managed to modify the ddr3 pcore for the roach2 to include the CPU > interface. I am working from a cloned ska-as mlib_devel repo. I sent the > changes/additions to Wesley New for vetting/committing to their repo but I > guess he hasn't done it yet. I can send the info to those interested. > Regards, mm. > > From: casper-boun...@lists.berkeley.edu > <mailto:casper-boun...@lists.berkeley.edu> [casper-boun...@lists.berkeley.edu > <mailto:casper-boun...@lists.berkeley.edu>] on behalf of Jack Hickish > [jackhick...@gmail.com <mailto:jackhick...@gmail.com>] > Sent: 15 December 2016 12:32 PM > To: Marc Welz; Franco > Cc: Casper Lists > Subject: Re: [casper] progdev fail > > > More broadly speaking, has anyone ever used the DDR powerPC interface on > ROACH2? Is it actually know to work with code as-provided in the mlib_devel / > katcp repos? > Franco -- perhaps this project -- > https://github.com/argonnexraydetector/RoachFirmPy > <https://github.com/argonnexraydetector/RoachFirmPy><https://github.com/argonnexraydetector/RoachFirmPy > <https://github.com/argonnexraydetector/RoachFirmPy>> -- which I found > linked on the DDR page of the CASPER wiki, might be interesting to you. > > Cheers > Jack > > > > > > On Thu, 15 Dec 2016 at 07:31 Marc Welz <m...@ska.ac.za > <mailto:m...@ska.ac.za><mailto:m...@ska.ac.za <mailto:m...@ska.ac.za>>> wrote: > The fpga<->ppc interface is limited in size - if you want to see huge > memory areas, you will need some sort of windowing/offset scheme > > regards > > marc > > > On Wed, Dec 14, 2016 at 1:43 PM, Franco <francocuro...@gmail.com > <mailto:francocuro...@gmail.com><mailto:francocuro...@gmail.com > <mailto:francocuro...@gmail.com>>> wrote: > > Doing some more testing, I realize the progdev works when I remove the DRAM > > block from the model. What could this mean?? > > > > Franco > > > > > > > > On 14/12/16 10:13, Franco wrote: > >> > >> Dear Casperites, > >> > >> I'm trying to progdev a model to ROACH2 and I get the following error: > >> > >> ?progdev cntr_dram2_2016_Dec_13_1558.bof > >> #log info 1019840342915 raw > >> attempting\_to\_program\_cntr_dram2_2016_Dec_13_1558.bof > >> #log info 1019840343002 raw > >> attempting\_to\_program\_bitstream\_of\_19586188\_bytes\_to\_device\_/dev/roach/config > >> #fpga loaded > >> #log warn 1019840343887 raw > >> requesting\_to\_map\_a\_rather\_large\_area\_of\_0x8000 > >> #log error 1019840343887 raw > >> unable\_to\_map\_file\_/dev/roach/mem:\_Cannot\_allocate\_memory > >> #log error 1019840343890 raw > >> unable\_to\_program\_bit\_stream\_from\_cntr_dram2_2016_Dec_13_1558.bof > >> !progdev fail > >> > >> Anyone has any idea what's the problem? I can program other models fine, > >> but this is the first model I compiled using the SKA fork library. > >> > >> Thanks, > >> > >> Franco > >> > > > > > > > > Please consider the environment before printing this e-mail > > > View the Reutech Radar System online disclaimer at > http://www.rrs.co.za/links/E-maildisclaimer.asp > <http://www.rrs.co.za/links/E-maildisclaimer.asp><116121517112301343.gif><116121517112301343.gif>
Re: [casper] progdev fail
Hi Mike, That would be great, thanks! Jack On Thu, 15 Dec 2016 at 15:11 Mike Movius <mi...@reutech.co.za> wrote: > > Hi, > I have managed to modify the ddr3 pcore for the roach2 to include the CPU > interface. I am working from a cloned ska-as mlib_devel repo. I sent the > changes/additions to Wesley New for vetting/committing to their repo but I > guess he hasn't done it yet. I can send the info to those interested. > Regards, mm. > > From: casper-boun...@lists.berkeley.edu [casper-boun...@lists.berkeley.edu] > on behalf of Jack Hickish [jackhick...@gmail.com] > Sent: 15 December 2016 12:32 PM > To: Marc Welz; Franco > Cc: Casper Lists > Subject: Re: [casper] progdev fail > > > More broadly speaking, has anyone ever used the DDR powerPC interface on > ROACH2? Is it actually know to work with code as-provided in the mlib_devel > / katcp repos? > Franco -- perhaps this project -- > https://github.com/argonnexraydetector/RoachFirmPy< > https://github.com/argonnexraydetector/RoachFirmPy> -- which I found > linked on the DDR page of the CASPER wiki, might be interesting to you. > > Cheers > Jack > > > > > > On Thu, 15 Dec 2016 at 07:31 Marc Welz <m...@ska.ac.za m...@ska.ac.za>> wrote: > The fpga<->ppc interface is limited in size - if you want to see huge > memory areas, you will need some sort of windowing/offset scheme > > regards > > marc > > > On Wed, Dec 14, 2016 at 1:43 PM, Franco <francocuro...@gmail.com francocuro...@gmail.com>> wrote: > > Doing some more testing, I realize the progdev works when I remove the > DRAM > > block from the model. What could this mean?? > > > > Franco > > > > > > > > On 14/12/16 10:13, Franco wrote: > >> > >> Dear Casperites, > >> > >> I'm trying to progdev a model to ROACH2 and I get the following error: > >> > >> ?progdev cntr_dram2_2016_Dec_13_1558.bof > >> #log info 1019840342915 raw > >> attempting\_to\_program\_cntr_dram2_2016_Dec_13_1558.bof > >> #log info 1019840343002 raw > >> > attempting\_to\_program\_bitstream\_of\_19586188\_bytes\_to\_device\_/dev/roach/config > >> #fpga loaded > >> #log warn 1019840343887 raw > >> requesting\_to\_map\_a\_rather\_large\_area\_of\_0x8000 > >> #log error 1019840343887 raw > >> unable\_to\_map\_file\_/dev/roach/mem:\_Cannot\_allocate\_memory > >> #log error 1019840343890 raw > >> unable\_to\_program\_bit\_stream\_from\_cntr_dram2_2016_Dec_13_1558.bof > >> !progdev fail > >> > >> Anyone has any idea what's the problem? I can program other models fine, > >> but this is the first model I compiled using the SKA fork library. > >> > >> Thanks, > >> > >> Franco > >> > > > > > > > Please consider the environment before printing this e-mail > > View the Reutech Radar System online disclaimer at > http://www.rrs.co.za/links/E-maildisclaimer.asp >
Re: [casper] progdev fail
Hi, I have managed to modify the ddr3 pcore for the roach2 to include the CPU interface. I am working from a cloned ska-as mlib_devel repo. I sent the changes/additions to Wesley New for vetting/committing to their repo but I guess he hasn't done it yet. I can send the info to those interested. Regards, mm. From: casper-boun...@lists.berkeley.edu [casper-boun...@lists.berkeley.edu] on behalf of Jack Hickish [jackhick...@gmail.com] Sent: 15 December 2016 12:32 PM To: Marc Welz; Franco Cc: Casper Lists Subject: Re: [casper] progdev fail More broadly speaking, has anyone ever used the DDR powerPC interface on ROACH2? Is it actually know to work with code as-provided in the mlib_devel / katcp repos? Franco -- perhaps this project -- https://github.com/argonnexraydetector/RoachFirmPy<https://github.com/argonnexraydetector/RoachFirmPy> -- which I found linked on the DDR page of the CASPER wiki, might be interesting to you. Cheers Jack On Thu, 15 Dec 2016 at 07:31 Marc Welz <m...@ska.ac.za<mailto:m...@ska.ac.za>> wrote: The fpga<->ppc interface is limited in size - if you want to see huge memory areas, you will need some sort of windowing/offset scheme regards marc On Wed, Dec 14, 2016 at 1:43 PM, Franco <francocuro...@gmail.com<mailto:francocuro...@gmail.com>> wrote: > Doing some more testing, I realize the progdev works when I remove the DRAM > block from the model. What could this mean?? > > Franco > > > > On 14/12/16 10:13, Franco wrote: >> >> Dear Casperites, >> >> I'm trying to progdev a model to ROACH2 and I get the following error: >> >> ?progdev cntr_dram2_2016_Dec_13_1558.bof >> #log info 1019840342915 raw >> attempting\_to\_program\_cntr_dram2_2016_Dec_13_1558.bof >> #log info 1019840343002 raw >> attempting\_to\_program\_bitstream\_of\_19586188\_bytes\_to\_device\_/dev/roach/config >> #fpga loaded >> #log warn 1019840343887 raw >> requesting\_to\_map\_a\_rather\_large\_area\_of\_0x8000 >> #log error 1019840343887 raw >> unable\_to\_map\_file\_/dev/roach/mem:\_Cannot\_allocate\_memory >> #log error 1019840343890 raw >> unable\_to\_program\_bit\_stream\_from\_cntr_dram2_2016_Dec_13_1558.bof >> !progdev fail >> >> Anyone has any idea what's the problem? I can program other models fine, >> but this is the first model I compiled using the SKA fork library. >> >> Thanks, >> >> Franco >> > > Please consider the environment before printing this e-mail. DISCLAIMER: This message is intended for the use of the addressee and may contain information that is privileged and/or confidential. If you are not the intended recipient, you are hereby notified that any dissemination, distribution or copying of the information contained in this message is strictly unauthorized and prohibited. If you have received this message in error, please notify the sender by reply e-mail and delete the message from your system. Opinions, conclusions, or other statements in this message that do not relate to the business of REUTECH Radar Systems, a Division of Reutech Ltd., its subsidiaries or affiliates, are neither given nor endorsed by REUTECH Radar Systems, a Division of Reutech (Pty)Ltd.
Re: [casper] progdev fail
Thanks Jack, I'll check into the project but it seems to have what I've been looking for. Franco On 15/12/16 07:28, Jack Hickish wrote: More broadly speaking, has anyone ever used the DDR powerPC interface on ROACH2? Is it actually know to work with code as-provided in the mlib_devel / katcp repos? Franco -- perhaps this project -- https://github.com/argonnexraydetector/RoachFirmPy -- which I found linked on the DDR page of the CASPER wiki, might be interesting to you. Cheers Jack On Thu, 15 Dec 2016 at 07:31 Marc Welz> wrote: The fpga<->ppc interface is limited in size - if you want to see huge memory areas, you will need some sort of windowing/offset scheme regards marc On Wed, Dec 14, 2016 at 1:43 PM, Franco > wrote: > Doing some more testing, I realize the progdev works when I remove the DRAM > block from the model. What could this mean?? > > Franco > > > > On 14/12/16 10:13, Franco wrote: >> >> Dear Casperites, >> >> I'm trying to progdev a model to ROACH2 and I get the following error: >> >> ?progdev cntr_dram2_2016_Dec_13_1558.bof >> #log info 1019840342915 raw >> attempting\_to\_program\_cntr_dram2_2016_Dec_13_1558.bof >> #log info 1019840343002 raw >> attempting\_to\_program\_bitstream\_of\_19586188\_bytes\_to\_device\_/dev/roach/config >> #fpga loaded >> #log warn 1019840343887 raw >> requesting\_to\_map\_a\_rather\_large\_area\_of\_0x8000 >> #log error 1019840343887 raw >> unable\_to\_map\_file\_/dev/roach/mem:\_Cannot\_allocate\_memory >> #log error 1019840343890 raw >> unable\_to\_program\_bit\_stream\_from\_cntr_dram2_2016_Dec_13_1558.bof >> !progdev fail >> >> Anyone has any idea what's the problem? I can program other models fine, >> but this is the first model I compiled using the SKA fork library. >> >> Thanks, >> >> Franco >> > >
Re: [casper] progdev fail
More broadly speaking, has anyone ever used the DDR powerPC interface on ROACH2? Is it actually know to work with code as-provided in the mlib_devel / katcp repos? Franco -- perhaps this project -- https://github.com/argonnexraydetector/RoachFirmPy -- which I found linked on the DDR page of the CASPER wiki, might be interesting to you. Cheers Jack On Thu, 15 Dec 2016 at 07:31 Marc Welzwrote: > The fpga<->ppc interface is limited in size - if you want to see huge > memory areas, you will need some sort of windowing/offset scheme > > regards > > marc > > > On Wed, Dec 14, 2016 at 1:43 PM, Franco wrote: > > Doing some more testing, I realize the progdev works when I remove the > DRAM > > block from the model. What could this mean?? > > > > Franco > > > > > > > > On 14/12/16 10:13, Franco wrote: > >> > >> Dear Casperites, > >> > >> I'm trying to progdev a model to ROACH2 and I get the following error: > >> > >> ?progdev cntr_dram2_2016_Dec_13_1558.bof > >> #log info 1019840342915 raw > >> attempting\_to\_program\_cntr_dram2_2016_Dec_13_1558.bof > >> #log info 1019840343002 raw > >> > attempting\_to\_program\_bitstream\_of\_19586188\_bytes\_to\_device\_/dev/roach/config > >> #fpga loaded > >> #log warn 1019840343887 raw > >> requesting\_to\_map\_a\_rather\_large\_area\_of\_0x8000 > >> #log error 1019840343887 raw > >> unable\_to\_map\_file\_/dev/roach/mem:\_Cannot\_allocate\_memory > >> #log error 1019840343890 raw > >> unable\_to\_program\_bit\_stream\_from\_cntr_dram2_2016_Dec_13_1558.bof > >> !progdev fail > >> > >> Anyone has any idea what's the problem? I can program other models fine, > >> but this is the first model I compiled using the SKA fork library. > >> > >> Thanks, > >> > >> Franco > >> > > > > > >
Re: [casper] progdev fail
The fpga<->ppc interface is limited in size - if you want to see huge memory areas, you will need some sort of windowing/offset scheme regards marc On Wed, Dec 14, 2016 at 1:43 PM, Francowrote: > Doing some more testing, I realize the progdev works when I remove the DRAM > block from the model. What could this mean?? > > Franco > > > > On 14/12/16 10:13, Franco wrote: >> >> Dear Casperites, >> >> I'm trying to progdev a model to ROACH2 and I get the following error: >> >> ?progdev cntr_dram2_2016_Dec_13_1558.bof >> #log info 1019840342915 raw >> attempting\_to\_program\_cntr_dram2_2016_Dec_13_1558.bof >> #log info 1019840343002 raw >> attempting\_to\_program\_bitstream\_of\_19586188\_bytes\_to\_device\_/dev/roach/config >> #fpga loaded >> #log warn 1019840343887 raw >> requesting\_to\_map\_a\_rather\_large\_area\_of\_0x8000 >> #log error 1019840343887 raw >> unable\_to\_map\_file\_/dev/roach/mem:\_Cannot\_allocate\_memory >> #log error 1019840343890 raw >> unable\_to\_program\_bit\_stream\_from\_cntr_dram2_2016_Dec_13_1558.bof >> !progdev fail >> >> Anyone has any idea what's the problem? I can program other models fine, >> but this is the first model I compiled using the SKA fork library. >> >> Thanks, >> >> Franco >> > >
Re: [casper] progdev fail
Doing some more testing, I realize the progdev works when I remove the DRAM block from the model. What could this mean?? Franco On 14/12/16 10:13, Franco wrote: Dear Casperites, I'm trying to progdev a model to ROACH2 and I get the following error: ?progdev cntr_dram2_2016_Dec_13_1558.bof #log info 1019840342915 raw attempting\_to\_program\_cntr_dram2_2016_Dec_13_1558.bof #log info 1019840343002 raw attempting\_to\_program\_bitstream\_of\_19586188\_bytes\_to\_device\_/dev/roach/config #fpga loaded #log warn 1019840343887 raw requesting\_to\_map\_a\_rather\_large\_area\_of\_0x8000 #log error 1019840343887 raw unable\_to\_map\_file\_/dev/roach/mem:\_Cannot\_allocate\_memory #log error 1019840343890 raw unable\_to\_program\_bit\_stream\_from\_cntr_dram2_2016_Dec_13_1558.bof !progdev fail Anyone has any idea what's the problem? I can program other models fine, but this is the first model I compiled using the SKA fork library. Thanks, Franco